SP2014 / Android-Malware-Detector

A machine learning based Android malware detection model.

Project Title

Android Malware Detection using Support Vector Machine Classifier

Prerequisites

• androguard - Reverse Engineering tool for android applications
• sklearn - Machine Learning library for python
• mongodb - A No-SQL database
• pymongo - Python driver for MongoDB

Instructions

1. Navigate to 'files' folder and place malware and benign apps into their respective folder.
2. Run 'python add_db.py' to save all the features extracted from all apks.
3. Run 'create_data.py' to get data from mongodb and save it to a csv file.
4. Once the 'data.csv' file is created, Run 'python feature_selection.py ' to perform feature selection, this will create two files namely 'final_data.csv' with optimal features and 'final_selected_features.txt' which contains the selected features along with their value.
5. Run 'python model_train.py <train_test/k-fold>' to train the model, a dump of the trained model is created.
6. Finally run 'python predict.py ' to perform the classification of that given apk.

Happy Coding.